Pressuposições e a Análise de Variância de Experimentos Agropecuários em Software Livre

Abstract

This paper provides a brief discussion of the history of modern statistics, taking into account the contributions of Fisher (1945). Focuses on, discuss the emergence of ANOVA (analysis of variance - ANOVA) and multiple comparison method, and at the same time, talk about the verification of the assumptions of the mathematical models in a completely randomized design and in blocks. The verification of assumptions will be through commands that trigger additional package features free software "R version 3.2.3." data from one experiment were used "tomato seedling production (Solanum lycopersicum) on different substrates" for discussion. The same was conducted at the Federal Institute Goiano - Morrinhos Campus, and the design adopted was randomized (DIC) with seven treatments and five replications. Data were summarized as means and arranged in an Excel table, being saved in format (.txt) that is readable by R. The commands were written in the editor notepad and pasted into the R prompt to verify the normality and homogeneity independence of errors. The R version 3.2.3 software has shown to be very efficient for the analysis of assumptions and analysis in ANOVA with subsequent multiple comparisons by Tukey test at 5% probability. Thus, the R software can be adopted in other work because of the quality of its results and its gratuitousness.